Predicting Site Response Predicting Site Response Based on

Predicting Site Response

Predicting Site Response • Based on theoretical calculations – 1 -D equivalent linear, non-linear – 2 -D and 3 -D non-linear • Needs geotechnical site properties

Imaging of Near-Surface Seismic Slowness (Velocity) and Damping Ratios (Q)

Image What? • Sβ(z) (shear-wave slowness) (=1/velocity) • Sα(z) (compressional-wave slowness) • ξβ(z) (shear-wave damping ratio [Qβ]) Why? • • Site amplification Site classification for building codes Identification of liquefaction and landslide potential Correlation of various properties (e. g. , geologic units and Vs)

Why Slowness? • Travel time in layers directly proportional to slowness; travel time fundamental in site response (e. g. , T = 4*s*h = 4*travel time) • Can average slowness from several profiles depth-by-depth • Slowness is the usual regression coefficient in fits of travel time vs. depth • Visual comparisons of slowness profiles more meaningful for site response than velocity profiles

Why Show Slowness Rather Than Velocity? Large apparent differences in velocity in deeper layers (usually higher velocity) become less important in plots of slowness Focus attention on what contributes most to travel time in the layers

Imaging Slowness • Invasive Methods – Active sources – Passive sources • Noninvasive Methods – Active sources – Passive sources

Invasive Methods • Active Sources – surface source – downhole source • Passive sources – Recordings of earthquake waves in boreholes---not covered in this talk

Invasive Method Surface Source-Downhole Receiver (ssdhr) (receiver can be on SCPT rod) One receiver moved up or down hole

SURFACE SOURCE ---SUBSURFACE RECEIVERS • downhole profiling – – velocities from surface data gaps filled by average velocity expensive (requires hole) depth range limited (but good to > 250 m) • seismic cone penetrometer – – advantages of downhole inexpensive limited range not good for cobbly materials, rock

Create a record section—opposite directions of surface source (red, blue traces) Pick arrivals (black) Plotting sideways makes it easier to see slopes changes by viewing obliquely (an exploration geophysics trick) CCOC

Finer layering in upper 100 m

Two models from the same travel time picks.

The increased resolution makes little difference in site amplification

SUBSURFACE SOURCE --- SUBSURFACE RECEIVERS • crosshole – “point” measurements in depth – expensive (2 holes) – velocity not appropriate for site response • suspension logger – – – rapid collection of data (no casing required) average velocity over small depth ranges can be used in deep holes expensive (requires borehole) no way of interpolating across data gaps

Downhole source--- P-S suspension logging (aka “PS Log”) Dominant frequency = 1000 Hz From Geovision

Example from Coyote Creek: note 1) overall trend; 2) “scatter”; 3) results averaged over various depth intervals reduces “noise”

“Noise” fluctuations in both S and P logs agree with variations in lithology! (No averaging)

Some Strengths of Invasive Methods • Direct measure of velocity • Surface source produces a model from the surface, with depth intervals of poor or missing data replaced by average layer (good for site amplification calculations) • PS suspension logging rapid, can be done soon after hole drilled, no casing required, not limited in depth range

Some Weaknesses of Invasive Methods • Expensive! (If need to drill hole) • Surface source may have difficulties in deep holes, requires cased holes, logging must wait • PS suspension log does not produce model from the surface (but generally gets to within 1 to 2 m), and there is no way of interpolating across depth intervals with missing data.

Noninvasive Methods • Active Sources – e. g. , SASW and MASW • Passive sources (usually microtremors) – Single station – Arrays (e. g. , fk, SPAC) • Combined active—passive sources

Overview of SASW and MASW Method • Spectral-Analysis-of-Surface-Waves (SASW— 2 receivers); Multichannel Analysis of Surface Waves (MASW— multiple receivers) • Noninvasive and Nondestructive • Based on Dispersive Characteristics of Rayleigh Waves in a Layered Medium

SASW Field Procedure • Transient or Continuous Sources (use several per site) • Receiver Geometry Considerations: – Near Field Effects – Attenuation – Expanding Receiver Spread – Lateral Variability (Brown)

SASW & MASW Data Interpretation Dispersion curve built from a number of subsets (different source, different receiver spreads) (Brown)

Some Factors That Influence Accuracy of SASW & MASW Testing • Lateral Variability of Subsurface • Shear-Wave Velocity Gradient and Contrasts • Values of Poisson’s Ratio Assumed in the inversion of the dispersion curves • Background Information on Site Geology Improves the Models

Noninvasive Methods • Passive sources (usually microtremors) – Single station (much work has been done on this method---e. g. , SESAME project. I only mention it in passing, using some slides from an ancient paper)

Ellipticity (H/V) as a function of frequency depends on earth structure (Boore & Toksöz, 1969)

Noninvasive Methods • Passive sources (usually microtremors) – Multiple stations (usually twodimensional arrays)

The array of stations at WSP used by Hartzell (Hartzell, 2005)

Inverting to obtain velocity profile (Hartzell, 2005)

Noninvasive Methods • Often active sources are limited in depth (hard to generate lowfrequency motions) • Station spacing used in passive source experiments often too large for resolution of near-surface slowness • Solution: Combined active—passive sources

An example from the CCOC—WSP experiment (active: f > 4 Hz; passive: f<8 Hz) (Yoon and Rix, 2005)

Comparing Different Imaging Results at the Same Site • Direct comparison of slowness profiles • Site amplification – From empirical prediction equations – Theoretical • Full resonance • Simplified (Square-root impedance)

Comparison of slowness profiles:

Coyote Creek Blind Interpretation Experiment (Asten and Boore, 2005) CCOC = Coyote Creek Outdoor Classroom

The Experiment • Measurements and interpretations done voluntarily by many groups • Interpretations “blind” to other results • Interpretations sent to D. Boore • Workshop held in May, 2004 to compare results • Open-File report published in 2005 (containing a summary by Asten & Boore and individual reports from participants)

Active sources at WSP: note larger near-surface & smaller deep slownesses than reference for most methods.

Passive sources at WSP: note larger near-surface & smaller deep slownesses than reference for most methods. Models extend to greater depth than do the models from active sources

Combined active & passive sources at WSP: note larger near-surface slownesses than reference

leading to these small differences in empirically-based amplifications based on V 30 (red=active; blue=passive & combined)

Average slownesses tend to converge near 30 m (coincidence? ) with systematic differences shallower and deeper (both types of source give larger shallow slowness; at 30 m the slowness from active sources is larger than the reference and on average is smaller than the reference for passive sources.

But larger differences at higher frequencies (up to 40%) (V 30 corresponds to ~ 2 Hz)

Summary (short) • Many methods available for imaging seismic slowness • Noninvasive methods work well, with some suggestions of systematic departures from borehole methods • Several measures of site amplification show little sensitivity to the differences in models (on the order of factors of 1. 4 or less) • Site amplifications show trends with V 30, but the remaining scatter in observed ground motions is large